Low-pH-Induced Lamellar to Bicontinuous Primitive Cubic Phase

Article Cite This: Langmuir XXXX, XXX, XXX-XXX

pubs.acs.org/Langmuir

Low-pH-Induced Lamellar to Bicontinuous Primitive Cubic Phase Transition in Dioleoylphosphatidylserine/Monoolein Membranes Toshihiko Oka,†,‡ Moynul Hasan,§ Md. Zahidul Islam,§,∥ Md. Moniruzzaman,§ and Masahito Yamazaki*,†,‡,§ †

Nanomaterials Research Division, Research Institute of Electronics, ‡Department of Physics, Graduate School of Science, and Integrated Bioscience Section, Graduate School of Science and Technology, Shizuoka University, Shizuoka 422-8529, Japan

§

S Supporting Information *

ABSTRACT: Electrostatic interactions (EIs) play important roles in the structure and stability of inverse bicontinuous cubic (QII) phases of lipid membranes. We examined the effect of pH on the phase of dioleoylphosphatidylserine (DOPS)/ monoolein (MO) membranes at low ionic strengths using small-angle X-ray scattering (SAXS). We found that the phase transitions from lamellar liquid-crystalline (Lα) to primitive cubic (QIIP) phases in DOPS/MO (2/8 molar ratio) membranes occurred in buffers containing 50 mM NaCl at and below the final pH of 2.75 as the pH of the membrane suspension was decreased from a neutral value. The kinetic pathway of this transition was revealed using time-resolved SAXS with a stopped-flow apparatus. The first step is a rapid transition from the Lα phase to the hexagonal II (HII) phase, and the second step is a slow transition from the HII phase to the QIIP phase. We determined the rate constants of the first step, k1, and of the second step, k2, by analyzing the time course of SAXS intensities quantitatively. The k1 value increased with temperature. The analysis of this result provided the values of its apparent activation energy, which were constant over temperature but increased with pH. This can be explained by an EI effect on the free energy of the transition state. In contrast, the k2 value decreased with temperature, indicating that the true activation energy increased with temperature. These experimental results were analyzed using the theory of the activation energy of phase transitions of lipid membranes when the free energy of the transition state depends on temperature. On the basis of these results, we discussed the mechanism of this phase transition. acids of peptides18,34 are used. Interactions of ions (such as H+ and Ca2+) and positively charged peptides also induced these phase transitions by changing the surface charge density of the membranes due to their binding to the charged groups.11,19,20,22,26,32 Generally, to understand the characteristics of various phase transitions of lipid membranes and their mechanisms, information about their kinetic pathways is indispensable.35−41 The first experimental evidence of the kinetic pathway of the EI-induced transition between the Lα and QII phase in lipid membranes has been revealed recently using time-resolved small-angle X-ray scattering (TR-SAXS); for the Lα to doublediamond QIID (or Q224) phase transition in a dioleoylphosphatidylserine (DOPS)/monoolein (MO) membrane induced by a decrease in pH, the first step is the direct transition from the Lα phase to a transient intermediate, an inverse hexagonal II (HII) phase, and the second step is the slow transition from the HII to QIID phase.42,43 The rate constants of the two steps were

1. INTRODUCTION Electrostatic interaction is one of the most important physical interactions in biological molecular systems, especially biomembranes such as plasma membranes and lipid membranes.1−5 Typically, biomembranes and lipid membranes are planar, lamellar membranes, which are in the liquid-crystalline (Lα) phase, but under certain conditions, they form regular three-dimensional (3D) structure membranes in inverse bicontinuous cubic (QII) phases, in which water channels with nanometer diameters were connected to each other.6−10 Recently, it has been well established that electrostatic interactions (EIs) due to surface charges of lipid membranes play various important roles in the structures and phase stabilities of the QII phases of biological lipid membranes. One of the EI effects on the QII phases is that the lattice constants and the size of water channels of the QII phase increase with an increase in the level of EI.11−15 The other EI effect is that the modulation of EI due to surface charges of membranes induces transitions between the Lα and QII phases and transitions between different QII phases in various lipids.11,12,16−34 As the surface charges, not only charged groups of hydrophilic segments of lipids but also charged side chains of amino © XXXX American Chemical Society

Received: July 19, 2017 Revised: September 28, 2017 Published: October 2, 2017 A

DOI: 10.1021/acs.langmuir.7b02512 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

prepare DOPS/MO multilamellar vesicles (MLVs), we added 100 μL of 10 mM ammonium acetate buffer (pH 6.7) containing 50 mM NaCl (buffer A) to a dry MO/DOPS lipid film (10 μmol)43 and then mixed the suspension several times using a vortex mixer for ∼20 s at room temperature (∼25 °C). To purify MLVs, we centrifuged the MLV suspension at 13000g for 20 min at 25 °C and then resuspended the pellet gently in new buffer A without using the vortex mixer. This suspension was used as the purified MLV suspension. We used the purified MLVs within 12 h of preparation. We determined the lipid concentrations of the suspensions using the Bartlett method.48 To measure the pH of DOPS/MO suspensions, a pH meter (HM-41X, DKK-TOA Co., Tokyo, Japan) was used, and the standard deviations of all the measured pH values were 5 h at ∼25 °C before the SAXS measurement. After incubation of the sample in the capillary in the holder at 25 °C for 10 min, the temperature of the sample was increased from 25 to 50 °C at a rate of 1 K/min. During the temperature scan, the SAXS patterns were measured with a time resolution of 5 s. Lattice constants were determined by fitting two Gauss functions to (110) and (200) peaks of the QIIP phase and one to the (10) peak of the HII phase. Peaks at larger angles were excluded from the fitting because their peak intensities were low. The positions of the peaks at larger angles were used in the phase determination of the samples. We used the same method of TR-SAXS in our previous paper44 to monitor structural changes in DOPS/MO membranes after the pH jump. TR-SAXS data were obtained at the BL40B2 beamline at SPring-8 (Sayo, Japan). A stopped-flow apparatus (SFM-CD10, UNISOKU, Osaka, Japan) was used to mix rapidly the purified DOPS/MO MLV suspension in buffer A with 20 mM citrate buffer at various pH values containing 50 mM NaCl (buffer C) at a volume ratio of 1/9. Other experimental conditions were described previously.44 The lattice constant, a, of the QIIP phase is determined by the equation S = (1/a)(h2 + k2 + l2)1/2, and that of the HII phase, which equals the center−center distance of adjacent cylinders in the HII phase, is determined by the equation S = [2/(31/2a)](h2 + k2 + hk)1/2, where S is the reciprocal spacing and h, k, and l are Miller indices. We used the non-negative matrix factorization (NMF) method based on a modified alternating least-squares (MALS) algorithm49,50 to obtain the time course of each phase in the SAXS patterns. The detailed method of analysis using the NMF method was described previously.44

determined by the analysis of the time course of SAXS peak intensities.43 The temperature dependence of the rate constants of this transition was also reported;44 the rate constants of the first step and those of the second step at pH 2.7 and 2.8 increased with temperature, their analysis provided the apparent activation energies, but the rate constant at the second step at pH 2.6 decreased with temperature, suggesting its activation energy varies with temperature. These results were analyzed using the quantitative theory of the activation energy of phase transitions of lipid membranes, which was obtained by modification of the theory of Squires et al.,45 including the effect of monolayer curvature. This new theory indicates that the free energy of the transition state depends on temperature and hence explains the results reasonably.44 Generally, the activation energy of structural changes can provide valuable information about their elementary processes and mechanisms. One of the determinants of the activation energy is the free energy of the transition state. The temperature can change the free energy of the transition state of the phase transitions involved in the QII phases in lipid membranes.44,45 Currently, we have a hypothesis that the EI may control the free energy of the transition state.44 In this report, to elucidate the EI effects on elementary processes of the Lα to QII phase transition in a DOPS/MO (2/8 molar ratio) membrane induced by a decrease in pH, we investigated this phase transition in the buffer containing 50 mM NaCl and also the temperature dependence of this phase transition. Debye lengths, measures of the effective length of electrostatic interaction, in a solution containing 50 and 100 mM NaCl are 1.4 and 0.96 nm, respectively.3 Therefore, we can reasonably expect that the EI in 50 mM NaCl is stronger than that in 100 mM NaCl, which was used in our previous investigation of the Lα to QII phase transition in a DOPS/MO (2/8) membrane induced by a decrease in pH.42−44 On the other hand, we demonstrated that the EI-induced transition between the Lα and QII phase in lipid membranes occurs mainly due to the change in the spontaneous curvature of monolayer lipid membranes.12,20,21,42,44 It is generally recognized that shortrange forces such as interactions between headgroups of neighboring lipids strongly affect the spontaneous curvature of monolayers, which was verified by experimental results.18,19,46,47 Therefore, we can reasonably consider that the electrostatic interactions in the presence of 100 and 50 mM NaCl have strong effects on the spontaneous curvature of the monolayer. Furthermore, we consider that this stronger EI in 50 mM NaCl may more strongly affect the transition states of this phase transition. In this work, we first found that in the presence of 50 mM NaCl a transition from the Lα phase to the primitive QIIP (or Q229) phase, not the QIID phase, occurred as the pH of the membrane suspension was decreased from a neutral value. Next we investigated the kinetics of this transition using TR-SAXS with a stopped-flow apparatus and found that the Lα phase was directly transformed into the HII phase during the first step and subsequently the HII phase slowly converted into the QIIP phase. By analyzing these results, we determined the rate constants of the two elementary steps. Finally, we investigated the temperature dependence of this phase transition and obtained information about the activation energy of the two elementary steps.

3. RESULTS 3.1. Effect of pH on the Structure of DOPS/MO Membranes in 50 mM NaCl. First, we examined the effect of pH on the structure of preformed DOPS/MO MLVs (2/8) in excess buffer containing 50 mM NaCl at 25 °C. The SAXS pattern of 5.0 mM purified MLVs of this membrane in ammonium acetate buffer (pH 6.7) showed a set of peaks with a spacing of 12 nm in a ratio of 1:2 (Figure 1a), indicating that it was in the Lα phase. We mixed various pH citrate buffers with the purified MLV suspension, and then the suspensions were incubated for 8−9 h before SAXS measurement. The final pH values of the suspensions were measured and are indicated in the texts and Figure 1. The structures of these samples greatly depended on the final pH of the suspension. For example, at a final pH of 2.55, the SAXS peaks that have a √2:√4:√6:√10:√12:√14:√16:√18 spacing ratio were

2. MATERIALS AND METHODS DOPS was purchased from Avanti Polar Lipids (Alabaster, AL). MO was purchased from Sigma-Aldrich Chemical Co. (St. Louis, MO). To B

DOI: 10.1021/acs.langmuir.7b02512 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

higher NaCl concentrations, and a transition between the QIIP and the QIID phases occurred at 75 mM NaCl (Figure 1d). 3.2. Kinetics of the Lα to QIIP Phase Transition Induced by Decreasing the pH. First, we examined the kinetic pathway of the Lα to QIIP phase transition in DOPS/MO (2/8) membranes in the presence of 50 mM NaCl at 25 °C, which was induced by decreasing the pH of the suspension from neutral pH to pH 2.55 (final pH). After mixing a purified MLV suspension in buffer A with buffer C (pH 2.40) using the stopped-flow apparatus, we continuously monitored the structure of the membranes in the suspension (final lipid concentration of 5.3 mM, final pH of 2.55). Figure 2a shows the time course of its SAXS patterns. The SAXS patterns from 0.2 to 10 s corresponded to the Lα phase due to MLV structure

Figure 1. Dependence of the structure and phase of a DOPS/MO (2/ 8) membrane on pH and salt concentration. SAXS patterns of a DOPS/MO (2/8) membrane suspension (a) in buffer A (pH 6.7) containing 50 mM NaCl and (b) in mixed buffer at pH 2.55 containing 50 mM NaCl after a long incubation time of 10 h. (c) pH dependence of the lattice constant of the QIIP phase in a DOPS/MO (2/8) membrane in low-pH buffer containing 50 mM NaCl. (d) Dependence of the lattice constant and phase in a DOPS/MO (2/8) membrane on NaCl concentration at pH 2.6: (●) QIIP phase and (■) QIID phase.

indexed as (110), (200), (211), (310), (222), (321), (400), and (330) (411) peaks of the QIIP phase, a bicontinuous, bodycentered cubic phase;12 its lattice constant was 15.2 nm (Figure 1b). Figure 1c shows the pH dependence of the structure of DOPS/MO (2/8) membranes. At and below pH 2.75, they were in the QIIP phase, and its lattice constant increased with pH. We also examined the effect of NaCl concentration on the phase and structure of DOPS/MO (2/8) membranes at a final pH of 2.6. The membranes were in the QIIP phase in the presence of low NaCl concentrations but in the QIID phase at

Figure 2. Changes in the structure and phase of DOPS/MO (2/8) membranes in 50 mM NaCl after the change in pH from pH 6.7 to a low pH at 25 °C. (a) SAXS contour plot of the time course at a final pH of 2.55. (b) Averaged SAXS patterns of the data shown in panel a from 0.2 to 9.2 s (bottom), from 50 to 89 s (middle), and from 2560 to 3629 s (top). (c) SAXS contour plot of the time course at a final pH of 2.65. C

DOI: 10.1021/acs.langmuir.7b02512 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir (Figure 2a,b). A new weak peak around S = 0.167 nm−1 was detected at ∼10 s, and its intensity grew over time. At 70 s, SAXS peaks with a spacing with a 1:√3:2 ratio appeared (Figure 2a,b), corresponding to the HII phase. Hence, we can consider that the peak at S = 0.167 nm−1 was the (10) peak of the HII phase. The peak intensities of the HII phase grew over time to ∼70 s and then decreased, while those of the Lα phase became low at ∼70 s. At ∼200 s, three weak peaks at 0.089, 0.125, and 0.154 nm−1 appeared, and their intensities grew over time (Figure 2a,b). The SAXS pattern of the same sample after 10 h was the same as that of Figure 1b, which corresponds to the QIIP phase as the equilibrium structure. Hence, we assigned the peaks at 0.089, 0.125, and 0.154 nm−1 in panels a and b of Figure 2 as the (110), (200), and (211) peaks of the QIIP phase, respectively. The analysis given above indicates that after the change in pH from a neutral value to a low value, the Lα phase converted into the HII phase rapidly and then the HII phase converted slowly into the QIIP phase. We also examined the pH dependence of the kinetics of this transition. The qualitative kinetic pathways of the transitions at all pH values were the same as that at pH 2.55. Figure 2c shows that at a final pH of 2.65 a peak due to the HII phase appeared ∼20 s after mixing, and then the peaks due to the QIIP phase appeared ∼200 s after mixing. The lattice parameters of the HII and QIID phases at pH 2.65 were larger than those at pH 2.55 (Table 1). Table 1. Temperature Dependence of the Lattice Constants at Various pH Values temperature (°C)

Lα(initial) (nm)

25 28 31

12 12 12

25 28 31

12 12 12

HII(intermediate) (nm)

QIIP(1 h) (nm)

Final pH of 2.55 7.0 6.9 6.9

15.9 15.3 14.9

7.1 7.0 6.9

17.3 16.4 15.7

Final pH of 2.65

3.3. Effect of Temperature on the Structure of the DOPS/MO (2/8) Membrane in 50 mM NaCl. We examined the temperature-induced transitions of DOPS/MO (2/8) membranes in their suspension (5.5 mM lipid) at various final pH values from 2.50 to 2.75. The contour plot shows the SAXS patterns of DOPS/MO (2/8) membranes at pH 2.55 recorded during a temperature scan from 25 to 50 °C at a rate of 1 K/min (Figure 3a). At 25 °C, the SAXS peaks that have a spacing ratio of √2:√4:√6 were indexed as (110), (200), and (211) peaks of the QIIP phase (Figure 3b). At 32 °C, a new weak peak appeared around 0.170 nm−1, and its intensity grew over time. At 50 °C, the peaks with a spacing ratio of 1:√3:2 appeared, which corresponds to the HII phase. Hence, the peak at 0.180 nm−1 was the (10) peak of the HII phase. Thereby, this result indicates that a QIIP to HII phase transition started to occur at 32 °C. At 50 °C, the peak intensity of the HII phase became strong, but weak peaks due to the QIIP remained around 0.117, 0.166, and 0.203 nm−1. At a final pH of 2.65, we obtained a similar result (Figure S1). The lattice constant of the QIIP phase was almost constant at low temperatures, but above 29 °C, it decreased with temperature (Figure 3c). On the other hand, the lattice

Figure 3. Changes in the structure and phase of DOPS/MO (2/8) membranes in 50 mM NaCl at low pH during a temperature scan from 25 to 50 °C at a rate of 1 °C/min. (a) SAXS contour plot of the temperature scan at a final pH of 2.55. (b) Averaged SAXS patterns of the data shown in panel a from 25.0 to 26.0 °C (bottom) and from 49.0 to 50.0 °C (top). (c) Temperature dependence of the lattice constants of the QIIP phase (top) and the HII phase (bottom) at a final pH of 2.55. (d) Temperature−pH phase diagram: (circles) QIIP phase and (triangles) coexistence of the QIIP and HII phases. D

DOI: 10.1021/acs.langmuir.7b02512 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

4. DISCUSSION In this report, we found that an Lα to QIIP phase transition occurred in DOPS/MO (2/8) membranes in buffers containing 50 mM NaCl at and below the final pH of 2.75 when the pH of the MLV suspension decreased from neutral pH (Figure 1c). As we reported previously,22,43 in the same buffer containing 100 mM NaCl, a different phase transition, i.e., an Lα to QIID phase transition, occurred in DOPS/MO (2/8) membranes. This difference can be explained as follows. The results of the salt concentration dependence of the membrane at pH 2.6 (Figure 1d) indicate that at c′, Ea(HII→QIIP) < 0. Under these conditions, ΔG⧧(HII→QIIP) increases greatly with temperature, and as a result, the k2 value decreases with an increase in temperature. Figure 6b indicates that Ea(HII→QIIP) < 0 and | Ea(HII→QIIP)| increases with temperature, which agrees with the prediction of eq 3. As we discussed in our previous paper, ΔG⧧(Lα→HII) is proportional to Ncoop (see eq 10 in ref 44). We have not yet succeeded in determining the value of Ncoop in the process of the Lα to HII transition, and thus, we represented the values of Ea(Lα→HII) per mole. Therefore, the numerical values of Ea(Lα→HII) in themselves are not important, but the comparison of them can be used. The temperature range used for the determination of the apparent activation energy is very limited (i.e., from 25 to 31 °C), and therefore, the accuracy of their values is not so high [the significant figures of the apparent activation energy are only one or two; i.e., its relative error (fractional uncertainty) is 10−30% (Table 3)]. If we could increase the temperature range of the measurement and hence the number of data points, its significant figure would increase to two or three; i.e., its relative error is 1−5%. However, for the conclusion of this report, one or two significant figures is sufficient. The small temperature range of the measurement of the rate constant of the low-pHinduced phase transition is mainly due to the fact that the QII phase existed in a limited temperature range (from 25 to 32 °C) (Figure 3). We tried to measure the rate constant of phase transitions from 20 °C, but it turned out that the reproducibility of the rate constants was not good. This is likely due to hysteresis of the temperature change; in this

⎛ ∂ΔG⧧ ⎞ ⎡ ∂ ln k ⎤ ⧧ ⎟=c Ea(Lα→HII) = −R ⎢ ⎥ = ΔG − T ⎜ ⎣ ∂(1/T ) ⎦ ⎝ ∂T ⎠ (2)

Equation 2 indicates that Ea(Lα→HII) does not depend on temperature, which agrees with the experimental results. On the other hand, according to the theory,44 for the second step of G

DOI: 10.1021/acs.langmuir.7b02512 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir measurement, we prepared the samples at 25 °C and then decreased the temperature to 20 °C to measure phase transitions, but the involvement of the process in the decrease in temperature may induce the hysteresis probably due to its slow rate of transition. When we started to investigate the low-pH-induced transitions between different QII phases and between the QII phase and other phases (the HII and Lα phases) many years ago,11 we did not consider any applications. At that time, we found that in oleic acid (OA)/MO (1/9) membranes as the pH decreased from a neutral value, a QIIP to QIID phase transition occurred at pH 5.8 and then a QIID to HII phase transition occurred at pH 5.3.11 After that, we found that in DOPS/MO (2/8) membranes as the pH decreased from a neutral value, an Lα to QIID phase transition occurred at pH 2.9.22 This phase transition was reversible; i.e., as the pH increased from a low value, a QIID to Lα phase transition occurred.22 We also found that DOPS/MO (2/8) large unilamellar vesicles (LUVs) were transformed into the QIID phase as well as DOPS/MO (2/8) MLVs when the pH values of suspensions were decreased.22 Recently, several researchers found the applications of this principle of the low-pH-induced transitions between the QII and HII phases,26,32 indicating that they are useful in targeted drug delivery and in cancer treatment. In this report, we found that in DOPS/MO (2/8) membranes low-pH-induced Lα to QII phase transitions occurred at various ion concentrations (QIIP phase at low ion concentrations and QIID phase at higher ion concentrations). The pH values inducing these phase transitions were low (pH